ES2212739B1 - ANALYZING SYSTEM FOR THE DETECTION OF REDUCING AND OXIDIZING GASES IN A CARRIER GAS AND GAS SENSOR BASED ON METAL OXIDES OF SEMI-CONDUCTOR TYPE. - Google Patents

Abstract

Analyzer system for the detection of reducing and oxidizing gases in a carrier gas and gas sensor based on semiconductor type metal oxides. Analyzer system for the detection of reducing and oxidizing gases in a carrier gas, which is characterized by the fact that the detection means are sensors based on semiconductor metal oxides that work in the absence of oxygen and by the fact that said system it comprises means for connecting to a chamber containing said sensors, and also, due to the fact that the processing and control means comprise a system for recognizing said gases in real time, which provides a diagram in which the sensors are located and identified. measurements made on said carrier gas. Allows gas analysis in real time.

Description

Gas detection analyzer system reducers and oxidants in a carrier gas and gas based sensor in metal oxides of semiconductor type.

The present invention relates to a system analyzer for the detection of reducing and oxidizing gases and the real-time control of the flow quality of a carrier gas. It also refers to the use, in the absence of oxygen, of the gas sensors based on metal oxides of type semiconductor.

Background of the invention

The most common technique for the evaluation of quality of a carrier gas, such as carbon dioxide, includes the use of specific chromatography equipment, which they include different types of detectors to ensure the Sensitivity and selectivity of pollutant analysis usual present in carbon dioxide. These teams, in addition if they are expensive, they have the disadvantage that they do not allow continuous monitoring of the gas flow that is being used in the production. They only perform analysis of specific samples. This technique is commonly used in production centers to be able to evaluate the quality of dioxide obtained, but as it is expensive equipment, can hardly be installed in any plant carbon dioxide consumer, such as a bottling of carbonated beverages. An alternative is the taking of specific samples that can be sent to the laboratory corresponding for analysis. However, this system also allows continuous monitoring of gas flow, in addition to have not negligible costs.

Analyzer systems are known in the market for the analysis of the quality of carbon dioxide, which They comprise different specialized equipment such as;

\ sqbullet

compound analyzers sulphurous, usually based on piro systems luminescence,

\ sqbullet

hydrocarbon analyzers aromatic, based on PID systems (Photo Ionizatión Detector) with ultraviolet ray lamp,

\ sqbullet

hydrocarbon analyzers Total, based on FID systems (Flame Ionization Detector).

These analyzer systems present the inconvenient that they are expensive to be installed in plants carbon dioxide consumers, and also that neither allow real-time analysis of a continuous flow of carbon dioxide.

There is no low system on the market cost that is able to perform an analysis, even partial, of Carbon dioxide quality in real time.

None of the usual techniques used to assess the quality of carbon dioxide is based on utilization of sensors based on semiconductor metal oxides.

Gas sensors based on semiconductor type metal oxides for gas detection, such as toxic gases in the atmosphere. It's about sensors Simple, low cost and resistant.

Sensors based on metal oxides of type semiconductor, have been developed for gas detection reducers and oxidizers in the presence of pure air and therefore in presence of oxygen.

It is known that in the presence of pure air, the active material or semiconductor metal oxide (type n), heated to a temperature between 300ºC and 500ºC, adsorbs atmospheric oxygen until you reach an equilibrium situation. Adsorption process of an oxygen atom involves the capture of an electron from the metal oxide conduction band. Therefore, when a sensor is in the presence of pure air and in balance, It has a high electrical resistance, also called base resistance

It is known that if the sensor is exposed to the presence of a reducing gas, it will react with oxygen adsorbed, establishing a new equilibrium situation. In in this case the concentration of adsorbed oxygen atoms will be inferior to the one that existed in the presence of pure air and this will result in a greater number of electrons in the band of driving. The result is a very significant decrease in the sensor resistance This effect is reversible since the sensor can recover its base resistance, if exposed again to The presence of pure air.

In the presence of an oxidizing gas, a competition for adsorption sites between said gas and the oxygen. This translates into a situation of new equilibrium in the that the resistance of the sensor increases. This effect is again reversible.

It is known that the principle of operation of the type of sensors described is summarized in the fact that the conductance of these devices changes progressively with the changes that occur in the composition of the atmosphere.

Oxide based sensors are not known. semiconductor type metal that allow gas detection reducers and oxidants, in complete absence of oxygen, in a atmosphere or stream of carrier gas.

Description of the invention

The objective of the present invention is to solve the mentioned disadvantages, developing an analyzer system for the detection of reducing and oxidizing gases in a gas carrier, which evaluates the quality of carrier gas in real time, using gas sensors based on metal oxides of type semiconductor working in the absence of oxygen.

In accordance with this objective, the system analyzer of the present invention comprises a plurality of detection means, calibration means and processing means and control of the acquisition and recognition of data, and characterized by the fact that said gas detection means they are sensors based on semiconductor metal oxides that they work in the absence of oxygen, due to the fact that said system comprises means for connecting said carrier gas to a chamber of measure that contains these sensors, and also because of the fact that said processing and control means comprise a system of recognition of these gases in real time, which provides a diagram with delimited decision zones, in which e identify the measurements made on said carrier gas.

Thanks to its features, the system analyzer allows an analysis of the quality of a gas carrier, such as carbon dioxide, in real time. Be It is a low cost equipment that is applicable in plants gas consumers, such as carbon dioxide, in carbonated beverage bottling plants.

According to the invention, the system is characterized by the fact that said calibration means they comprise a plurality of calibrated standards or gases, so less equal to the number of reducing and oxidizing gases that must be detected in the carrier gas, and by the fact that the response of the plurality of sensors to the measurements of said patterns comprises obtaining a vector of conductance variation for each gas Calibrated or pattern.

According to the invention, said system of recognition includes obtaining a learning matrix resulting from grouping the conductance variation vectors of the measurements made with the plurality of patterns or gases calibrated

In accordance with the objective of this invention, said recognition system identifies the measures carried out in the carrier gas, according to the algorithm:

-

obtaining a conductance variation vector for the plurality of sensors that integrate the system.

-

self climbing vector with the means and variances used to autoscale the learning matrix obtained from patterns or gases calibrated

-

projection of autoscaled vector on the space of the main components extracted from the learning matrix obtained with the calibration means

-

In function of the position occupied by said vector, the system identifies a type of answer.

Preferably, the type of response identified by the system includes carrier gas responses pure, contaminated carrier gas at alert level of at least one pollutant and carrier gas contaminated at alarm level so Less a pollutant.

Advantageously, the system is characterized by the fact that said processing and control means comprise a microprocessor that corrects the temporal drifts of the responses of the sensors and controls and processes the data that allow to detect the presence of reducing and / or oxidizing gases at different levels Preset

Preferably, said connection means they comprise a plurality of solenoid valves and connecting tubes to allow carrier gas or calibrated gases to flow through of the camera containing the sensors.

According to a preferred embodiment of the invention, The carrier gas is carbon dioxide.

According to the invention, the use of a gas sensor based on metal oxides of type semiconductor for the detection of reducing and oxidizing gases present in a carrier gas, in the absence of oxygen.

The absence of oxygen in the carrier gas is refers to oxygen values in said gas, not exceeding 30 ppm of oxygen.

Brief description of the drawings

For greater understanding of how much has been exposed some drawings are accompanied in which, schematically and only to non-limiting example title, a case study of realization.

In these drawings,

Figure 1 shows the response of two sensors based on tin oxides, at 15 ppm of methane diluted in dioxide of carbon (Figure 1a) and 1 ppm of sulfur dioxide diluted in carbon dioxide (figure 1b).

Figure 2 shows a block diagram of operation of the analyzer system.

Figure 3 shows a diagram or space of the main components with delimited areas or decision areas for pure carbon dioxide, carbon dioxide at the level of Alert and carbon dioxide at alarm level.

Figure 4 shows a diagram or space of the main components in which three measures have been projected made in carbon dioxide, corresponding to the three vectors I1, I2 and I3.

Description of a preferred embodiment

Figures 1a and 1b show the answer of two gas sensors based on tin oxide (semiconductor type n), in the presence of traces of methane and sulfur dioxide, respectively, diluted in carbon dioxide.

Gas sensors based on metal oxides semiconductor type of the present invention detect gases oxidizers and reducers in a continuous stream of dioxide carbon, in the absence of oxygen. That is, at a concentration of oxygen in carbon dioxide at levels not exceeding 30 ppm of oxygen.

Figure 1a of the present invention shows the 1 response of a sensor, based on tin oxide (semiconductor of type n), at 15 ppm of methane diluted in carbon dioxide, as well as the resistance 2 base of the same sensor when it is in equilibrium situation with carbon dioxide. The presence of reducing gas leads to a decrease in sensor resistance 2. This effect is reversible since the sensor can recover its base resistance 2 if reducing gas is eliminated, in this case the methane, and is exposed to carbon dioxide again. Figure 1b of The present invention shows the response 3 of another sensor based in tin oxide (type n semiconductor) at 1 ppm of sulfur diluted in carbon dioxide, as well as the base resistance 4 of the same sensor when it is in equilibrium situation with carbon dioxide. In this case, the presence of an oxidizing gas leads to an increase in resistance 4 of the sensor. This effect It is reversible again.

It has been observed that carbon dioxide interacts reversibly with the surface of the metal oxide, acting in a similar way as oxygen does in detection of traces of pollutants (reducing and oxidizing gases).

Gas sensors in general and sensors based on metal oxides in particular, they are poorly selective. This It implies that all sensors have different responses, but not zero, to the polluting gases. Therefore it is necessary use a matrix of several sensors based on metal oxides, with partially overlapping sensitivities, in order to identify the different polluting gases in carbon dioxide.

Figure 2 of the present invention shows a block diagram that facilitates understanding of operation of the analyzer system. Said system consists of a camera of measure, stainless steel, containing seven 23 based sensors in metal oxides, equipped with different noble metals, catalytically active. The number of sensors is related to the number of pollutant gases that must be detected in the gas carrier 21 or carbon dioxide, whose quality you want to evaluate. In the example that concerns us the sensors have been chosen for the gas detection such as methane, propane, butane, hexane and others organic compounds, such as ethylene. The system comprises means of connection 22 of carbon dioxide 21 to the measuring chamber that it contains the sensors 23. It is a variable number of solenoid valves and connecting pipes, in stainless steel, for allow the gas whose quality you want to evaluate, or 26 different calibrated gases or calibration standards, can flow through the chamber where the sensors are located 23. The Gas flow should be set at a constant value, preferably 100 ml / min.

The resistance reading of the sensors is carried out by means of a semi-resistance bridge, where one resistance is the sensor itself ( Rs ) and the other is a fixed resistance of appropriate value ( Rm ). A known voltage ( Vc ) is applied to both resistors connected in series and the voltage at the intermediate point ( Vm ) is measured. This voltage is converted from analog to digital voltage, using an analog-to-digital A / D converter.

The heating of the sensors is carried out through electronic circuits that allow sensors to be heated up to its optimum operating temperature (between 300 and 400 ° C).

A program carried out by a microprocessor 24 is responsible for performing the following functions:

a) Control of solenoid valves both during the normal measurement phase as during the calibration phase of the equipment.

b) Control of the acquisition process of the Vm voltages and their A / D conversion.

c) Obtaining the voltage Vm for each sensor in the cluster, once per second, and calculating the average value of the Vm over a minute. Therefore, for each sensor, the average Vm values are calculated based on the last sixty measurements performed.

d) Obtaining the resistors for each of the sensors in the cluster, using the average values of the Vm . Thus, for the ith sensor of the grouping, knowing the average value Vm_ {i}, its resistance R_ {i} is obtained through the calculation:

R_ {i} = VcR_ {m} / (Vc-Vm_ {i})

e) Calculation of the conductance variations, once the resistance values of the sensors have been calculated. For the ith sensor, its conductance variation ΔG_ {i} is defined by the expression: ΔG_ {i} = 1 / R_ {i} - 1 / R_ {oi}, where R_ {oi} is the resistance of the sensor in the presence of the pure analyzed gas or base resistance of the ith sensor.

f) Obtaining the conductance variation vector for each measure; vector I = (\ DeltaG_ {1}, ..., \ DeltaG_ {n}), where n is the number of sensors that make up the cluster. Said vector constitutes the input information for a recognition algorithm that is responsible for evaluating the quality of the gas being analyzed.

g) Periodic download of the information generated by the analyzer system at an Internet address. Whether detect levels of some contaminant above values By default, the system can send alert and / or alarm messages to mobile phones.

The analyzer system described comprises a data recognition system 24 based on a process of learning that takes measurements using a set of patterns or calibrated gases 26. The response of the recognition system 24 It will be one of the following three:

to)

Identified pure carbon dioxide

b)

Identified polluted carbon dioxide, concentration in alert level. The pollutant (s) is (are): list of pollutants

C)

Identified polluted carbon dioxide, concentration in alarm level. The pollutant (s) is (are): list of pollutants

The learning process consists of taking measurements using a set of calibrated standards or gases 26. The patterns consist of pure carbon dioxide and contaminated carbon dioxide. Two calibrated standards are used for each contaminant considered: one standard with the contaminant diluted to the alert concentration in carbon dioxide and the other diluted to the alarm concentration. Finally, standards with binary mixtures of contaminants are also used. Each measure is repeated a minimum of three times to achieve a representative set of measures. Said learning process makes it possible to obtain a learning matrix, A , resulting from grouping the vectors of variation of conductance obtained in response to the learning measures described above. The dimension of A is m \ timesn , where m (rows) is the number of learning measures and n (columns) is the number of sensors that make up the grouping. Thus, each of the rows of A corresponds to one of the learning measures, and each of the columns of A contains the conductance variations experienced by a cluster sensor.

Matrix A is normalized by an autoscale operation. This operation is performed by columns. The mean and standard deviation of each column are obtained. If M_ {i} and D_ {i} are respectively the mean and standard deviation of column i , then the new value of any element of that column is calculated as new by e_ {ki} = (e_ {ki} - M_ {i}) / D_ {i}, where e_ {ki} represents the element in row k , column i in matrix A. Once A has been self-climbing, an extraction of main components is performed. The main component extraction process is a standard technique not described here. The main components are formed by a linear combination of the columns of the self-climbing matrix A.

The result of component extraction main performed on a set of measures with patterns or calibrated gases, is a diagram 25 of arbitrary units in the that these measures are located. The last step of the process learning consists in defining decision boundaries between areas of pure carbon dioxide, contaminated dioxide at the level of Alert and contaminated dioxide at alarm level.

Figure 3 shows a diagram 25 with zones or decision areas delimited for pure carbon dioxide, the carbon dioxide at alert level and carbon dioxide at alarm level Said diagram 25 has been obtained with the process of learning using the measurements with the calibrated gases or standards 26. Reference 00 corresponds to a measurement with dioxide of pure carbon, while references 01, 02, 03, 04, 05, 06, 14 and 32, correspond to carbon dioxide measurements with different types or mixtures of contaminants.

In figure 3:

00: pure CO2

01: CO2 + 10 ppm ethylene

02: CO2 + 20 ppm ethylene

03: CO2 + 30 ppm methane

04: CO2 + 15 ppm methane

05: CO2 + 0.5 ppm dioxide sulfur

06: CO2 + 1 ppm dioxide sulfur

14: CO2 + 10 ppm ethylene + 15 ppm methane

32: CO2 + 20 ppm ethylene + 30 ppm methane

In said diagram 25 it can be seen as the pure carbon dioxide clearly differs from the dioxide of polluted carbon The different types of pollution, since measures of the same type, for example pure carbon dioxide, they appear grouped in the space of the two first major components and carbon dioxide measurements affected by different pollutants, occupy different positions in the space of the first two main components.

The data recognition system 24 described identifies the measurements made on carrier gas 21, in this Carbon dioxide case, according to the algorithm:

to)

Obtaining a new conductance variation vector or new measure that is necessary to identify.

b)

Autoscale the vector, using the means and variances that were used to autoscale the learning matrix A.

C)

Projection of autoscaled vector on component space main.

d)

In function of the position occupied by said vector, the system decides what kind of output is the one identified.

Figure 4 shows a diagram or space 25 of the main components in which three have been projected measures, corresponding to the three vectors I1, I2 and I_ {3}. These measures must be identified by the system of recognition 24. Given the position occupied by the projection I_ {1}, it is concluded that said measure belongs to pure carbon dioxide. The position of I_ {2} is very close to that of the measures of methane calibration, it is concluded that this measure corresponds at an alert level for methane contamination in the dioxide of carbon. The position of I_ {3} is very close to that of the measures calibration with methane and ethylene, it is concluded that this measure is corresponds to an alarm level for methane contamination and ethylene in carbon dioxide.

Gas sensors 23 suffer drifts Temporary in your response. These drifts are associable to processes of aging of the active material. Therefore to keep smooth operation of the analyzer system over time, It is necessary to perform calibrations periodically. The system makes these calibrations in an automated and transparent way to the user / operator. With a predetermined periodicity, for example Every twenty-four hours, the system enters autocalibration mode. In this mode the following steps are performed:

to)

Performing a measurement with each of the calibrated standards: pure carbon dioxide, carbon dioxide with contaminant alarm level 1, carbon dioxide with contaminant alarm level p . Where p is the number of contaminants that are detected in carbon dioxide.

b)

Autoscaling and projection of calibration measures over space 25 of the main components.

C)

Computation of the distances of each of the calibration measurements to the centroid of the class they belong to. If this distance exceeds a certain preset level, new decision boundaries are recalculated taking take into account the information provided by the measures of calibration.

Once the self-calibration process is finished, the team is ready to continue with the analysis in real time of the quality of carbon dioxide or carrier gas.

Although it has been described and represented a concrete embodiment of the present invention, it is evident that the subject matter expert may introduce variants and modifications, or replace the details with technically equivalent ones, without depart from the scope of protection defined by the claims attached.

Claims (9)

1. Analyzer system for the detection of reducing and oxidizing gases in a carrier gas 21, comprising a plurality of detection means 23, calibration means 26 and processing and control means 24 of data acquisition and recognition, characterized by the the fact that said gas detection means are sensors 23 based on semiconductor metal oxides that work in the absence of oxygen, due to the fact that said system comprises connection means 22 of said carrier gas 21 to a measuring chamber containing said sensors 23, and in addition, by the fact that said processing and control means 24 comprise a real-time recognition system of said gases, which provides a diagram 25 with delimited decision zones, in which the sensors are located and identified. measurements carried out on said carrier gas 21. 2. System according to claim 1 characterized in that said calibration means comprise a plurality of calibrated standards or gases 26, at least equal to the number of reducing and oxidizing gases that must be detected in the carrier gas 21, and by the fact that the response of the plurality of sensors 23 to the measurements of standards 26 comprises obtaining a conductance variation vector for each calibrated gas or standard 26. 3. System according to claims 1 and 2, characterized in that said recognition system comprises obtaining a learning matrix resulting from grouping the conductance variation vectors of the measurements made with the plurality of calibrated standards or gases 26 . 4. System according to claim 3, characterized in that said recognition system identifies the measurements made on the carrier gas 21, according to the algorithm:

-

obtaining a conductance variation vector for the plurality of 23 sensors that integrate the system.

-

self climbing vector with the means and variances used to autoscale the learning matrix obtained from patterns or gases calibrated 26.

-

projection of autoscaled vector on the space 25 of the components main extracted from the learning matrix obtained with calibration means 26.

-

In function of the position occupied by said vector, the system identifies a type of answer.

5. System according to claim 4, characterized in that the type of response identified by the system comprises the responses of pure carrier gas, contaminated carrier gas at the alert level of at least one contaminant and contaminated carrier gas at the level of alarm of at least one contaminant. System according to claim 1, characterized in that said processing and control means 24 comprise a microprocessor that corrects the temporal drifts of the responses of the sensors 23 and controls and processes the data that allow the presence of reducing gases to be detected. and / or oxidants at different preset levels. 7. Analyzer system according to claims 1 and 2, characterized in that said connecting means 22 comprise a plurality of solenoid valves and connecting tubes to allow carrier gas 21 or calibrated gases 26 to flow through the chamber containing the sensors 23. 8. Analyzer system according to claim 1, characterized in that the carrier gas 21 is carbon dioxide. 9. Use of a gas sensor 23 based on semiconductor type metal oxides for gas detection reducers and oxidants present in a carrier gas 21, in the absence of oxygen